Deterministic AST engine for AI coding agents. We map codebase architecture into a local Rust graph so your LLM stops guessing. Sub-ms latency. Zero token burn.
Joined April 2026
- Tweets 116
- Following 73
- Followers 38
- Likes 11
78 Photos and videos
Jun 8
Hi @garrytan, could you please confirm whether the Paxel client is open source? Iβm interested in doing a security review before using it in practice.
74
Jun 3
Three questions every CTO should be able to answer in 2 minutes,
file:line:
1. Where do PII fields flow through my codebase?
2. Which SSRF/SQLi/auth-bypass sinks are reachable from untrusted input?
3. Did my AI agent's new handler traverse the rate limiter?
Most teams can't answer in 2 days.
1
58
Jun 1
Claude Opus 4.8 scores ~88% on SWE-bench Verified. Going higher is brutal.
We connected it to ArgosBrain β an engine that instantly looks up any codebase and gives the model exact answers β and pushed it to 91.4%.
22 hard tasks it failed alone. Now solved. π
π github.com/CataDef/LongMemCoβ¦
2
92
May 15
I just published The Re-Read Tax: 73% of Claude Code Tokens Are Redundant medium.com/p/the-re-read-taxβ¦
1
161
May 15
Weekly leaderboard time.
Drop how many tokens you burned this week coding with Claude / Cursor / Copilot / Codex.
By Sunday I'm crowning the King of Token Burners
(prize: bragging rights the unique pleasure of explaining to your CFO where they all went)
I'll start: 22M tokens on Claude Code this week.
GO. Your number?
1
155
May 11
π Heading to Paris May 21 to demo at @aitinkerers Paris β
"From 73% Token Waste to Zero"
I'll reveal the full architecture of Kubernetes (~2M LOC) live β every symbol, every call graph, every cross-file reference exposed as a queryable graph in seconds.
Then watch a Claude Code agent on it drop per-task cost 10Γ in real time. No slides. Pure Rust MCP.
If you're building agent infra, see you there π¦
paris.aitinkerers.org/p/ai-tβ¦
92
May 5
Uber's CTO confirmed publicly: Entire 2026 AI budget burned in 4 months That's a $85β114M annual run-rate on Claude Code alone. Anthropic's own /context data: 73% of a Claude Code bill is the agent re-reading code it already read. ArgosBrain replaces that with a sub-millisecond graph lookup. 0 tokens per query. Reduces ~70% of the repo-re-read chunk. Math: $43β83M/year ArgosBrain would have eliminated from Uber's bill. Same agents. Same code output. Smaller invoice. argosbrain.com β Aurelian Jibleanu
1
2
118
May 3
Your AI coding agent forgets your codebase every single session.
β re-reads the same files 40x/week
β burns ~$1,350/mo on grep read summarize loops
β hallucinates symbols that don't exist
ArgosBrain gives it a brain.
Deterministic graph memory. 0 tokens per query. P99 sub-ms.
Drop-in MCP for @AnthropicAI Claude Code, @cursor_ai, Codex, Cline, Aider.
Local. Free tier.
argosbrain.com
cc @simonw @swyx @alexalbert__ @mckaywrigley
#AI #AIAgents #ClaudeCode #Cursor #MCP #DevTools
2
94